[GitHub] [hudi] nsivabalan commented on a change in pull request #4971: [HUDI-3556] Re-use rollback instant for rolling back of clustering and compaction if rollback failed mid-way

[GitBox]

nsivabalan commented on a change in pull request #4971:
URL: https://github.com/apache/hudi/pull/4971#discussion_r824330180



##########
File path: 
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/client/BaseHoodieWriteClient.java
##########
@@ -954,29 +954,42 @@ private HoodieTimeline 
getInflightTimelineExcludeCompactionAndClustering(HoodieT
     return inflightTimelineExcludeClusteringCommit;
   }
 
-  private Option<HoodiePendingRollbackInfo> 
getPendingRollbackInfo(HoodieTableMetaClient metaClient, String 
commitToRollback) {
-    return getPendingRollbackInfos(metaClient).getOrDefault(commitToRollback, 
Option.empty());
+  protected Option<HoodiePendingRollbackInfo> 
getPendingRollbackInfo(HoodieTableMetaClient metaClient, String 
commitToRollback) {
+    return getPendingRollbackInfo(metaClient, commitToRollback, true);
+  }
+
+  protected Option<HoodiePendingRollbackInfo> 
getPendingRollbackInfo(HoodieTableMetaClient metaClient, String 
commitToRollback, boolean ignoreCompactionAndClusteringInstants) {
+    return getPendingRollbackInfos(metaClient, 
ignoreCompactionAndClusteringInstants).getOrDefault(commitToRollback, 
Option.empty());
+  }
+
+  protected Map<String, Option<HoodiePendingRollbackInfo>> 
getPendingRollbackInfos(HoodieTableMetaClient metaClient) {
+    return getPendingRollbackInfos(metaClient, true);
   }
 
   /**
    * Fetch map of pending commits to be rolled-back to {@link 
HoodiePendingRollbackInfo}.
    * @param metaClient instance of {@link HoodieTableMetaClient} to use.
    * @return map of pending commits to be rolled-back instants to Rollback 
Instant and Rollback plan Pair.
    */
-  protected Map<String, Option<HoodiePendingRollbackInfo>> 
getPendingRollbackInfos(HoodieTableMetaClient metaClient) {
+  protected Map<String, Option<HoodiePendingRollbackInfo>> 
getPendingRollbackInfos(HoodieTableMetaClient metaClient, boolean 
ignoreCompactionAndClusteringInstants) {
     List<HoodieInstant> instants = 
metaClient.getActiveTimeline().filterPendingRollbackTimeline().getInstants().collect(Collectors.toList());
     Map<String, Option<HoodiePendingRollbackInfo>> infoMap = new HashMap<>();
     for (HoodieInstant instant : instants) {
       try {
         HoodieRollbackPlan rollbackPlan = 
RollbackUtils.getRollbackPlan(metaClient, instant);
         String action = rollbackPlan.getInstantToRollback().getAction();
-        if (!HoodieTimeline.COMPACTION_ACTION.equals(action)) {
-          boolean isClustering = 
HoodieTimeline.REPLACE_COMMIT_ACTION.equals(action)
-              && ClusteringUtils.getClusteringPlan(metaClient, 
instant).isPresent();
-          if (!isClustering) {
-            String instantToRollback = 
rollbackPlan.getInstantToRollback().getCommitTime();
-            infoMap.putIfAbsent(instantToRollback, Option.of(new 
HoodiePendingRollbackInfo(instant, rollbackPlan)));
+        if (ignoreCompactionAndClusteringInstants) {
+          if (!HoodieTimeline.COMPACTION_ACTION.equals(action)) {
+            boolean isClustering = 
HoodieTimeline.REPLACE_COMMIT_ACTION.equals(action)
+                && ClusteringUtils.getClusteringPlan(metaClient, new 
HoodieInstant(true, rollbackPlan.getInstantToRollback().getAction(),
+                
rollbackPlan.getInstantToRollback().getCommitTime())).isPresent();
+            if (!isClustering) {
+              String instantToRollback = 
rollbackPlan.getInstantToRollback().getCommitTime();
+              infoMap.putIfAbsent(instantToRollback, Option.of(new 
HoodiePendingRollbackInfo(instant, rollbackPlan)));
+            }
           }
+        } else {
+          
infoMap.putIfAbsent(rollbackPlan.getInstantToRollback().getCommitTime(), 
Option.of(new HoodiePendingRollbackInfo(instant, rollbackPlan)));

Review comment:
       synced up offline. resolving the comment.




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscr...@hudi.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org